simtb_match_EST2TRUE() - Matches estimated to true SMs based on maximum R^2 statistics
Usage:
>> [mIND, rIND, mSTRENGTH, Rvalues, RMAT] = simtb_match_EST2TRUE(SMest, SMtrue)
INPUTS:
SMest = [nEC x nVV] estimated SMs, where nEC is the number of
estimated components and nVV is the number
of voxels to analyze (may be masked data).
SMtrue = [nC x nVV] true SMs (saved in *_DATA.mat),
where nC is the true number of sources.
NOTE: SMest and SMtrue can be the mean (or aggregate) SMs across the group
or the SMs for a single subject. The number of estimated components (nEC)
may be greater or less than the number of true sources (nC).
OUTPUTS:
mIND = indices that best map SMtrue to SMest, i.e., SMest ~ SMtrue(mIND,:);
rIND = indices that best map SMest to SMtrue, i.e., SMtrue ~ SMest(rIND,:);
mSTRENGTH = strength of the match, defined as mSTRENGTH = (M1-M2)/(M1+M2), where
M1 is the maximum absolute beta coefficient, and M2 is the second greatest.
Betas are determined from the multiple linear regression where true components
are the predictors and the estimated component is the observation.
mSTRENGTH ranges from [0,1], 0 indicating a poor match, 1 indicating a good match.
Rvalues = R^2-statistics between the estimated and true components
RMAT = Full R2-statistic matrix, size [nEC x nC].
see also: simtb_regression()